Stainless steel has been widely applied in high-temperature environments containing water vapor, such as heat exchangers in power plants, boiler components and interconnectors in solid oxide fuel cells [1, 2, 3, 4, 5, 6, 7]. As it is well known, the oxidation performance of stainless steel mainly depends on the potential to form and maintain a Cr-rich protective oxide scale. However, in the presence of water vapor, the changes in the structure and performance of the protective oxide scale must be reconsidered, since many investigations have proved that water vapor could influence the high temperature oxidation process of stainless steel [7, 8]. The effects of water vapor on the oxidation of alloy, especially on stainless steel can be divided into:
It could promote the internal oxidation of Cr-based alloy ;
The conductivity of the formed scale could be increased in water vapor containing atmosphere ;
Essuman et al.  and Othman et al.  summarized the oxidation mechanisms of alloy in water vapor containing atmospheres. One of the important mechanisms is that water can modify the diffusion mode of oxygen in the oxide scale [16, 17]. This is because of the internal oxidation of metal under the oxide scale. Another important problem with respect to the oxidation of stainless steel in oxidizing atmospheres is Cr evaporation, especially in water vapor containing atmosphere. The analysis of the Cr-vapor species was also widely reported [18, 19, 20, 21, 22].
However, the oxidizing rate of stainless steel in low water vapor containing oxidizing atmosphere has not been widely examined in literature. Even the conclusions earlier reported are contradictory to many reports [10, 11, 23]. In the present work, the corrosion kinetics of the stainless steel 441, one of the typical materials using at high temperature, was studied by thermogravimetry in dry and wet air. The oxidation process and its mechanism of the stainless steel 441 in wet air are discussed.
The investigated stainless steel 441 was supplied by HuaTsing Power Sci&tech Co. The chemical compositions are shown in Table 1. The main elements were rechecked by the China National Center for Quality Supervision and Testing of Iron and Steel. The steel specimens, cut into sheets of 0.7×20×40 mm with specular gloss, were cleaned in alcohol, and dried before oxidation treatment. A small hole (diameter is 2 mm) was drilled in the upper part of the alloy so that a thin alumina rod could be inserted for hanging the specimen in the thermoanalyzer.
The specimens were oxidized at 1,000 °C for 8.3 h in air with high water vapor contents (>10 %) and 13.9 h with low water vapor contents (<10 %), respectively. This temperature is within a common oxidation test temperature range of 18 wt%-Cr-containing steel. Mass changes of the specimens during the oxidation process were recorded by the balance to an accuracy of 0.1 mg (made by Shanghai Youke instrument Co. Ltd.). Figure 1 shows the device which connects the specimen to the balance. The gas flow was maintained at 100 ml/min STP (accuracy ±1.8 ml/min) by means of an Alicat Gas Flow Controller. And liquid water was pumped into a steam generator (temperature was kept at 300 °C) using a MasterFlex peristaltic pump (accuracy ±0.1 %). Water vapor and air were mixed in the generator, and then fed into the furnace.
A scanning electron microscope (SEM, CARL ZEISS EVO MA 10/LS 10 JS) with energy dispersive spectrum (EDS, Thermo NORAN System) was employed to probe the topography of the specimens after oxidation. Raman spectrum (Horyba LabRAM HR Evolution) was also applied to investigate the component and growth stress of the oxidized specimens. The sample surface was probed with a 100 μm-spot-sized laser (532 nm laser in 25 mW).
Results and discussion
Mass-change of the specimens
Figure 2(a) is the mass-change profiles of the stainless steel 441 in wet atmosphere at 1,000 °C. The mass-change is expressed as Δm/S, where Δm is the mass gain of the specimen, and S is the exposure area of the specimen. The initial oxidation rate of the specimen was relatively fast in the air with low water vapor content (0 vol. % and 3.5 vol. % H2O). Usually, the mass gain profiles followed a quasi-parabolic rule. When the water vapor content was elevated to no less than 11.5 vol. %, the initial oxidation rate was lower. However, the oxidation slightly accelerated with the prolonged exposure time. It was due to the non-protective oxides scale and evaporation of Fe and Cr bearing substances, which will be introduced later. The oxidation kinetic profiles of Fe-Cr alloys in H2O containing atmosphere [11, 13] also showed the similar breakaway process. This process is usually described as following equation, (1)
where Δm is the sample weight change, A is the sample surface area, t is the oxidation time. The two constants, kd and ks, describe the diffusion process within the oxidized steel and Cr evaporation in the oxidation, respectively. Equation (1) indicates that the initial mass gain rate of alloy in moist high temperature atmosphere is the Highest during the process, and then it will decrease to a stable value. It can be noticed that the oxidation process of steel in dry atmosphere can be perfectly described by this model, where kd=2.09×10–5 mg·cm–2·s–1 and ks=4.51×10–6 mg·cm–2·s–1.
The oxidation of steel in moist atmosphere did not follow the quasi-parabolic mechanism. Further, the oxidation rate tended to be constant without any noticeable increase with prolonged exposure. In this situation, the process could be approximately treated as a simplified pattern of eq. (1): kd is assumed to be a relatively large number comparing to ks, in which it is implied that the protection of oxides scale is not such effective. It results in the eq. (1) to be (2)
where the physical meaning of k changes to be apparent reaction rate constant. Values of k of the mass-change profiles of the samples in various atmospheres were worked out for comparison. Considering the comparability, mass–gain curve in dry air was also fitted by eq. (2). Figure 2(b) shows that the absolute value of k decreased firstly, and then increased exponentially with increasing water vapor content in the atmosphere. Wet air with 3.5 vol. % H2O and dry air were beneficial to the oxidation resistance of stainless steel 441, which had the lowest average oxidation rate.
Topography of the oxide scale
With the increase of the water vapor content in the atmosphere, the topography of the oxide scale varied. Magnification views of the near surface region are shown in Figure 3. The oxide grains were fine in low water vapor-containing atmospheres. The grains sintered more seriously with the increase of water vapor content. When the water vapor content was as high as 11.5 vol. %, the oxide grains were compacted fully due to sintering. A large number of small “bubbles” and broken ones were found spreading all over the scale surface. The “bubbles” on the surface implied the formation of vapor molecules under the dense scale in the oxidation.
Big “bubbles” were found on the scale which developed in high water vapor containing atmospheres (Figure 4). The similar phenomenon has earlier been reported [24, 25, 26]. Surface of the “bubbles” was porous. Fe2O3 grains in the shape of blades were found on the surface of the “bubbles” (Figure 4(d)). The diameters of the “bubbles” were about ten to hundreds of micrometers. The “bubbles” expanded and were connected to each other with the increase of time and water vapor content. EDS investigations revealed that these consisted of (Fe,Cr) spinel and Fe2O3. Total covered area of the Fe-rich “bubbles” on the Cr2O3 scale was larger in air with higher water vapor content.
The oxides in the broken “bubbles” were analyzed. Figure 5 shows that the region was composed of Fe-rich oxides. It implied that a serious oxidization occurred on the broken zone in wet air for the non-protective from oxides scale.
The cross section of the oxidized stainless steel was shown in Figure 6. EDS results showed that the oxide layer contained mainly Cr and O. The thickness of the scales in dry air is about 4 μm, which is thicker than the scale thickness (1 μm) in wet air containing 3.5 vol. % water vapor. It indicated that the oxidation rate of the stainless steel 441 in wet air (3.5 vol. % H2O) was relatively slow. And the increased Fe content on the surface of oxide layer in Figure 6(a) also indicated that the oxidation of stainless steel 441 in dry air was more serious than it in air with 3.5 vol. % H2O. Oxides scales developed in wet air with more than 10 vol. % H2O shows two different cross sections. Figure 6(c) shows the most part of stainless steel 441 was covered with Cr2O3 based dense scale, which was determined by EDS. While Figure 6(d) shows that thick scale with Fe-rich oxides formed at some part. As mentioned, it was a part of expanded Fe-rich “bubble”. EDS results indicated that the oxides in Figure 6(d) from the surface of scale to the interface of oxides/steel were iron oxides, (Fe,Cr)3O4 and Cr2O3 in sequence.
Figure 7 shows the topography of the scale in wet air at 1,000 °C with different oxidation times. The scale which initially formed in 11.5 vol. % H2O containing air was dense and the boundaries of the grains were identifiable. With the exposure time increasing, some Cr2O3 fine grains developed from the dense Cr2O3 scale, and then sintered to be a uniform scale (Figure 3(d)).
The specimens were also analyzed by Raman spectra to confirm the oxidation products in two different stages, as shown in Figure 8. The bands around 225, 498 (A1g modes), 247, 293, 299, 412 and 613 cm–1 (Eg modes) were attributed to hematite . The Raman bands of Cr2O3 were around 235, 296.3, 350.4, 528.5, 615.0 (Eg modes), 266 and 554.2 cm–1 (A1g modes) [28, 29]. Bands around 664–685 cm–1 and 680–690 cm–1 were attributed to corundum solid solution and spinel phases respectively . The relative intensity of the band around 680–690 cm–1 (spinel phases) increased with the increase of exposure time. Meanwhile, the relative intensity of the band around 554.2 cm–1 (Cr2O3) decreased. It implied that the content of Cr2O3 was lower and the content of Fe-containing spinel phases was higher in the oxide layers when the exposure time was prolonged. Thus, the oxidation resistance of the oxide layer was weakening with the increased time in the oxidation process. But the main composition of the oxide layers was still Cr2O3 within 90 ks.
The Raman shift of the strongest band of Cr2O3 (554.2 cm–1) was recorded (as shown in Figure 9). The shifts of Raman bands were mainly due to the growth stress in the oxidation process . So the Raman shift of the band could be used to explore the growth stress of the oxidation process [29, 32]. The relaxed oxide growth on the stainless steel was reported at 549.3 cm–1 . Figure 8 shows that the Raman shift of the band increases with the increasing exposure time in a parabolic-like relationship. The band shift was not obvious after more than 60 ks of oxidation in wet air. The stresses were calculated by the equation in Ref. [28, 29, 32]: (3)
where Δν is the band shift to the reference wavenumber 549.3 cm–1, P is the growth stress of the scale. Thus, the influence of exposure time on the growth stress of the oxide scales could also be expressed as a parabolic-like relationship, which is similar to the oxidation process of 441 in dry oxygen. In the oxidation of Fe-18Cr-TiNb alloy, the growth stress of the scale increases quickly to around 2 GPa at about 3 ks, then decreases very slowly . But the growth stress of scale in wet air was found to increase within about 70 ks, which was much longer than the one in the dry oxygen. The slower increase of the growth stress implied that the stress caused by volume expansion of the formed oxide was getting relaxed. There might be two reasons for the relaxation. Firstly, the relative serious internal oxidation of the alloy eased the epitaxial stress of the oxide scales and the metal. Secondly, the formation and breakage of “bubbles” released part of the growth stress.
The oxidation acceleration of alloy in high water vapor containing atmosphere was reported by many earlier workers [10, 11, 23, 33]. However, it was concluded that the effects of water vapor on the stainless steel oxidation were complex, without a specific conclusion as to whether it could be beneficial or detrimental . The complex effects of water vapor could lead to various results in practice. The oxidation kinetics of the alloy at high temperature always could be quasi-parabolic or linear. In our study, the oxidation of stainless steel 441 in air with less than 3.5 vol. % H2O was following quasi-parabolic mechanism. When the water vapor content increased, the alloy was in “break away” oxidation. The oxidation rate of stainless steel samples increased with increasing time when the water vapor content was higher than 10 vol. %. Water vapor showed complex effects on the oxidation behavior of stainless steel.
The oxide scale structure of stainless steel 441 in water vapor containing atmosphere was different from that in dry air. Small content of water vapor promoted the sintering of oxide grains. It could be due to the transportation of Cr-containing vapor between the chromium oxide grains. With the existence of more water vapor, it would result in oxidative evaporation at high temperatures. At the same time, the sintering of the oxide grains was accelerated with the increasing water vapor content, which promoted the densification of the scale. This condition should be favorable toward resistance to further oxidation. Under this condition, the inward diffusion of oxygen, hydrogen and other particles could be more difficult than that in dry air atmosphere. As a result, the principal effect of water vapor on the parabolic oxidation process of the steel is likely to be the increase in diffusion resistance through the oxide scale in air with 3.5 vol. % H2O.
On the other hand, higher Cr-containing vapor at water vapor containing atmosphere would cause dilution of Cr in the scales [25, 33]. Figure 10 shows the thermodynamic diagram of two dominant species of chromium volatilization. As the figure shows, CrO3 is the main vapor product in the oxidation of Cr in dry air at high temperatures. When the atmosphere turns to be wet, CrO(OH)3 becomes the main type of vapor. It can be found that the vapor pressures of CrO3 and CrO(OH)3 increase with enhancing oxygen partial pressure . Based on the reported thermodynamic data [35, 36], the partial pressure of CrO2(OH)2 will exceed that of CrO3 when the water vapor pressure is above about 10–2 atm at 1,000 °C in air. The total vapor pressure of the products in the oxidation process of stainless steel increases with the increasing content of water vapor in high temperature atmosphere. Thus, it could be concluded that water vapor could result in high Cr-containing vapor pressures in the oxidation of stainless steel, and CrO2(OH)2 could have a strong impact on the oxidation process at high water vapor content.
While, Fe-containing vapors also acts important role in the oxidation process of stainless steel at high temperature. The observed blade iron oxides on the “bubbles” could be the precipitate of Fe-containing vapors. The observations of earlier researchers are similar to the present results, viz. blade hematite grains were observed on the oxide layers of stainless steel [24, 37, 38], which could be attributed to the vapor of the products . The loss of iron in iron-reached oxides layer resulted porous structure of it, which promoted the further break away oxidation of stainless steel.
The oxidation rate of stainless steel 441 in wet air with 3.5 vol. % H2O was found to be as low as that in dry air. Water vapor promoted the oxidation resistance of oxidized stainless steel due to the sintering of the oxide grains in low water vapor containing atmosphere.
The oxidation process of stainless steel 441 in dry air followed quasi-parabolic model. The rate constants kd and ks were 2.09×10–5 and 4.51×10–6 mg·cm–2·s–1, respectively.
In the oxidation process of stainless steel 441 in steam containing atmosphere, the growth stress of oxides scale increased relatively slowly, which increased to the maximum value within 70 ks. And a slow increasing content of Fe-containing spinel phase was found at the same time.
The increasing water vapor content resulted in numbers of defects in the scale. This, in turn, promoted Fe- and Cr-containing species evaporation. The vapors compared the breakage of the chromium oxide scale. Porous Fe-rich oxides layer developed from the “bubbles” and broken places, expanding on the scale, which caused the “break-away” oxidation of the steel in high water vapor containing atmosphere.
Y. Zhao and J. Fergus, J. Electrochem. Soc., 159 (2012) C109–C113. Google Scholar
M.P. Brady, M. Fayek, J.R. Keiser, H.M. Meyer Iii, K.L. More, L.M. Anovitz, D.J. Wesolowski and D.R. Cole, Corros. Sci., 53 (2011) 1633–1638. Google Scholar
Y. Larring, R. Haugsrud and T. Norby, J. Electrochem. Soc., 150 (2003) B374–B379. Google Scholar
H. Nickel, Y. Wouters, M. Thiele and W. Quadakkers, Fresenius J. Anal. Chem., 361 (1998) 540–544. Google Scholar
J. Shen, L. Zhou and T. Li, Oxid. Met., 48 (1997) 347–356. Google Scholar
M. Krumpelt, T.A. Cruse, B.J. Ingram, J.L. Routbort, S. Wang, P.A. Salvador and G. Chen, J. Electrochem. Soc., 157 (2010) B228–B233. Google Scholar
M. Halvarsson, J.E. Tang, H. Asteman, J.E. Svensson and L.G. Johansson, Corros. Sci., 48 (2006) 2014–2035. Google Scholar
J. Mougin, N. Rosman, G. Lucazeau and A. Galerie, J. Raman Spectro., 32 (2001) 739–744. Google Scholar
A. Srisrual, S. Coindeau, A. Galerie, J.P. Petit and Y. Wouters, Corros. Sci., 51 (2009) 562–568. Google Scholar
H. Asteman, J.E. Svensson and L.G. Johansson, Corros. Sci., 44 (2002) 2635. Google Scholar
J.W. Fergus, Mater. Sci. Eng. A., 397 (2005) 271–283. Google Scholar
W.-K. Burton, N. Cabrera and F. Frank, Philos. Trans. Roy. Soc. London A., (1951) 299–358. Crossref
About the article
Published Online: 2016-10-06
Published in Print: 2017-07-26
The authors gratefully thank the financial support of National Program on Key Basic Research Project of China (Grant No. 2012CB215405), China Postdoctoral Science Foundation (No. 2015M570036) and National Nature Science Foundation of China (No.51474141). The authors also thank Ms. Xiao-Jia Du for material supply.